Related Experiment Videos
Regional inference with averaged P values increases the power to detect linkage
L R Goldin1, G A Chase, A F Wilson
1Genetic Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, NIH, Bethesda, Maryland 20892-7236, USA. goldinl@exchange.nih.gov
Genetic Epidemiology
|August 14, 1999
Summary
Determining the best critical value for genomic linkage analysis is key. Averaging P values over 9-15 cM intervals, particularly using two consecutive loci, significantly improves linkage detection power compared to single-locus tests.
Area of Science:
- Genetics
- Statistical Genetics
- Genomic Analysis
Background:
- Establishing appropriate critical values for linkage analysis in genomic screening remains a challenge.
- Various critical values have been proposed for both single-locus and multi-locus linkage tests.
Purpose of the Study:
- To evaluate regional test criteria based on multiple single-locus analyses for linkage testing.
- To determine optimal critical values using simulation methods across different genetic map densities.
Main Methods:
- Simulations were conducted under the null hypothesis of no linkage.
- Tests evaluated included single loci, consecutive single loci, and moving averages of loci.
- Power comparisons were made between regional tests and standard single-locus tests.
Main Results:
- Averaging P values over 9-15 centimorgans (cM) intervals yielded the highest statistical power.
- Testing the average P value from two consecutive loci outperformed testing individual loci.
- The observed increase in power ranged from 7% to 29% across simulations.
Conclusions:
- Regional tests, specifically averaging P values over 9-15 cM intervals, offer superior power for linkage detection.
- Utilizing consecutive loci averages provides a more robust approach than single-locus testing in genomic screens.